chore: Update comment and ref for vllm #2527
Conversation
krishung5
requested review from
a team,
alec-flowers,
ishandhanani,
nnshah1,
ptarasiewiczNV,
richardhuo-nv,
rmccorm4 and
tanmayv25
as code owners
|
/ok to test d0cff4c |
WalkthroughUpdates vLLM references: Dockerfile comment bumped from v0.10.0 to v0.10.1; install script removes a comment line; pyproject optional dependency switches from pinned PyPI vllm==0.10.0 to a specific Git revision (77a6bf0...). No functional logic changes in scripts. Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Possibly related PRs
Poem
Tip 🔌 Remote MCP (Model Context Protocol) integration is now available!Pro plan users can now connect to remote MCP servers from the Integrations page. Connect with popular remote MCPs such as Notion and Linear to add more context to your reviews and chats. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. CodeRabbit Commands (Invoked using PR/Issue comments)Type Other keywords and placeholders
Status, Documentation and Community
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
pyproject.toml (1)
59-59: Switch to VCS pin looks fine; verify packaging/consumer impact and optionally use tagMoving vllm to a Git commit is acceptable, and you already have allow-direct-references = true configured for Hatch. Two follow-ups to avoid surprises:
- Developer installs of ai-dynamo[vllm] now require Git and build deps; this may fail on minimal environments that previously consumed PyPI wheels. Consider calling this out in docs or using a tag for self-documentation.
- If the intent is strictly “match v0.10.1,” you can pin by tag for clarity. Commit pinning is fine for reproducibility, but the tag makes intent obvious.
You can switch to the tag:
- { git = "https://github.com/vllm-project/vllm.git", rev = "77a6bf07aedf132aad2b6719f6d87abc5d3311ab" }, + { git = "https://github.com/vllm-project/vllm.git", rev = "v0.10.1" },Or keep the commit and add a comment for clarity:
"nixl<=0.4.1", - { git = "https://github.com/vllm-project/vllm.git", rev = "77a6bf07aedf132aad2b6719f6d87abc5d3311ab" }, + # vLLM v0.10.1 + { git = "https://github.com/vllm-project/vllm.git", rev = "77a6bf07aedf132aad2b6719f6d87abc5d3311ab" },To proactively check for packaging and consumer impact:
- Verify that pip/uv can resolve and install the extra in a clean environment (requires Git).
- Confirm the commit matches v0.10.1 so Dockerfile and pyproject remain in sync.
If you want, I can add a README note about the new requirement for Git and build tools when using the vllm extra.
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (3)
container/Dockerfile.vllm(1 hunks)container/deps/vllm/install_vllm.sh(0 hunks)pyproject.toml(1 hunks)
💤 Files with no reviewable changes (1)
- container/deps/vllm/install_vllm.sh
🧰 Additional context used
🧠 Learnings (1)
📚 Learning: 2025-07-22T10:22:28.972Z
Learnt from: ptarasiewiczNV
PR: ai-dynamo/dynamo#2027
File: container/deps/vllm/install_vllm.sh:0-0
Timestamp: 2025-07-22T10:22:28.972Z
Learning: The `--torch-backend=auto` flag works with vLLM installations via uv pip install, even though it's not a standard pip option. This flag is processed by vLLM's build system during installation to automatically match PyTorch distribution with container CUDA versions.
Applied to files:
container/Dockerfile.vllm
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (2)
- GitHub Check: Mirror Repository to GitLab
- GitHub Check: Build and Test - dynamo
| # Match 0.10.1 vLLM release | ||
| # https://github.com/vllm-project/vllm/releases/tag/v0.10.1 |
There was a problem hiding this comment.
💡 Verification agent
🧩 Analysis chain
Confirm commit/tag alignment and annotate for traceability
Good to document the target release. Since VLLM_REF (Line 16) is pinned to a commit, please verify that 77a6bf07aedf132aad2b6719f6d87abc5d3311ab is exactly the commit for v0.10.1, and consider annotating the commit in the comment for quicker traceability.
If you want to make this explicit, you could add a small comment:
# Match 0.10.1 vLLM release
# https://github.com/vllm-project/vllm/releases/tag/v0.10.1
+# VLLM_REF pinned to commit for v0.10.1
+# 77a6bf07aedf132aad2b6719f6d87abc5d3311abAlso, please sanity-check DEEPGEMM_REF and FLASHINF_REF compatibility with v0.10.1 (they are unchanged). If there are known pairings for this vLLM release, capturing them in a short comment helps future bumps.
🌐 Web query:
What commit hash does the vLLM GitHub tag v0.10.1 point to?
💡 Result:
The v0.10.1 tag points to commit aab549870df50edf0512f0a59b574f692f546465 (short: aab5498). (github.com)
Citations:
Align VLLM_REF with the v0.10.1 tag commit
The Dockerfile currently pins VLLM_REF to 77a6bf07aedf132aad2b6719f6d87abc5d3311ab, but the GitHub v0.10.1 tag actually points to commit aab549870df50edf0512f0a59b574f692f546465. Please update the ref and annotate it for traceability:
# Match 0.10.1 vLLM release
# https://github.com/vllm-project/vllm/releases/tag/v0.10.1
-# VLLM_REF pinned to commit for v0.10.1
-# 77a6bf07aedf132aad2b6719f6d87abc5d3311ab
+VLLM_REF=aab549870df50edf0512f0a59b574f692f546465
+# VLLM_REF pinned to commit for v0.10.1 (aab549870df50edf0512f0a59b574f692f546465)• After updating, please also sanity-check that your DEEPGEMM_REF and FLASHINF_REF remain compatible with this v0.10.1 commit.
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| # Match 0.10.1 vLLM release | |
| # https://github.com/vllm-project/vllm/releases/tag/v0.10.1 | |
| # Match 0.10.1 vLLM release | |
| # https://github.com/vllm-project/vllm/releases/tag/v0.10.1 | |
| VLLM_REF=aab549870df50edf0512f0a59b574f692f546465 | |
| # VLLM_REF pinned to commit for v0.10.1 (aab549870df50edf0512f0a59b574f692f546465) |
🤖 Prompt for AI Agents
In container/Dockerfile.vllm around lines 19 to 20, VLLM_REF is pinned to commit
77a6bf07aedf132aad2b6719f6d87abc5d3311ab but should point to the v0.10.1 tag
commit aab549870df50edf0512f0a59b574f692f546465; update the VLLM_REF value to
aab549870df50edf0512f0a59b574f692f546465 and add a brief comment explaining this
maps to GitHub v0.10.1 for traceability, then rebuild or re-run any local docker
steps to verify no build errors and sanity-check DEEPGEMM_REF and FLASHINF_REF
(update them if they are known incompatible) by ensuring their commits or tags
are compatible with v0.10.1 and adjusting pins or comments accordingly.
| "uvloop", | ||
| "nixl<=0.4.1", | ||
| "vllm==0.10.0", | ||
| { git = "https://github.com/vllm-project/vllm.git", rev = "77a6bf07aedf132aad2b6719f6d87abc5d3311ab" }, |
There was a problem hiding this comment.
we cannot use URLs in wheels dependencies, pypi wont allow us to publish python wheels which pull deps from URL
|
Closing. #2528 |
Overview:
Follow up PR for #2509.
Details:
Where should the reviewer start?
Related Issues: (use one of the action keywords Closes / Fixes / Resolves / Relates to)
Summary by CodeRabbit